Neuroparhology and Applied Neurobiology 199 1, 17,469-484 Coronavirus induced primary demyelination: indications for the involvement of a humoral immune response F. ZIMPRICH*, J. WINTER?, H. WEGET AND H. LASSMA*$ *Research Unit for Experiment Neuropathology, Austrian Academy o f Sciences, Vienna, Austria, tlnstitute for Virology and Immunology, University o f Wurzburg, Wurzburg, Germany and $Neurological Institute, University o f Vienna, Vienna, Austria ZIMPRICH F., WINTER J., WEGE H. Kyuwa Nagashima et al., 1978, 1979; Sorensen, Percy Wege et al., 1983; Koga, Wege Sorensen et al., 1984; Wege, Domes Watanabe, Wege Weiner, 1973), immune mediated destruction of virus infected cells or virus induced autoimmune reactions (Watanabe, Wege Wang, Stohlman Wege et al., 1984a). Animals Specific pathogen free Lewis rats were obtained from the Zentralinstitut fur Versuchstierzucht, Hannover, FRG. The rats were intracerebrally inoculated with 800 PFU of the MHV-JHM virus, between 3 and 8 weeks of age. Histology Animals were perfused with 4% paraformaldehyde dissolved in 0.1 M phosphate buffer. Parts of the brain and spinal cord were embedded in paraffin. Paraffin sections (3 pm thick) were stained with haematoxylin Vass et al., 1989). Biotinylated species specific anti-mouse and MHV-JHM induced demyelination and humoral immune response 47 1 anti-rabbit sera were purchased from Amersham, UK; horse radish peroxidase labelled Avidin from Sigma, St Louis, USA, rabbit anti-mouse and APAAP complex from Dakopats, Denmark. The following primary antibodies and sera were used: a polyclonal rabbit anti JHM serum (Wege, Watanabe mouse monoclonal antibodies against nucleo- capsid (N 556) and surface protein (S-D15 and S-E16) (Wege et ul., 1984a); a polyclonal rabbit serum against complement component 9 (Linington et al., 1989); a rabbit serum against rat albumin (Nordic, Netherlands) and a biotinylated anti rat IgG serum (Amersham, UK) to stain plasma cells; monoclonal antibodies against glial fibrillary acidic protein (GFAP) (Boeringer Mannheim, Germany); myelin basic protein MBP (Hybtritec, USA); myelin oligodendroglial glycoprotein (MOG) (8-18-C5) and PO (a gift from Dr C. Linington). ED1 (Serotec, UK) and W3/13 (Sera-lab, UK) were applied to stain monocytes, T cells and neutrophils respectively. For identification of the infected glial cell type (oligodendrocytes or astrocytes) we used both a morphological approach (as shown in Figure 5b,c) and a double labelling immunohisto- chemistry procedure. As a first step viral antigen was visualized using the polyclonal rabbit serum and the avidin/biotin technique. Subsequently, oligodendrocytes and astrocytes were identified on the same sections using mouse monoclonal antibodies against myelin oligo- dendroglial glycoprotein (MOG) (8- 18-C5) or glial fibrillary acidic protein (GFAP) and the alkaline phosphatase anti-alkaline phosphatase (APAAP) technique. Immune electron microscopy For immune electron microscopy an avidin/biotin technique was employed (Lassmann et al., 1986). Sections cut with a razor blade were incubated for 8 h in each step with the following antisera, diluted in phosphate-buffered saline containing 10% fetal calf serum (FCS). Monoclonal antibodies against nucleocapsid (N556) or spike protein (S-D 15 and S-E 16) were used as a primary layer; sections were then incubated with biotinylated anti-mouse immuno- globulin (Amersham, UK), to which 3% rat serum was added to abolish cross reactivity completely. Horse radish peroxidase labelled avidin (Sigma, St Louis, USA) was applied in the next step and the reaction was visualized by a diaminobenzidine reagent. Sections were osmicated and embedded in Epon. Quantitative evaluation Quantitative studies on virus antigens were performed in the spinal cord white matter of 14 rats from different stages of the disease. N as well as S protein expressing glial cells were evaluated in consecutive sections at least in two spinal cord levels per animal. Similarly plasma cells (IgG+), macrophages (ED1 +) and Tcells (W3/13+) were counted in at least two spinal cord levels of all animals in each group. Non-parametric Mann-Whitney U test was used to analyse statistical differences in the infiltration with inflammatory cells and in the relative proportion of S expressing cells between AE, CPE and SDE. RESULTS Clinical disease The following study was performed with tissue samples from 29 diseased Lewis rats which were selected for detailed neuropathological investigation from a total number of 400 infected Lewis 472 F. Zimprich et al. Table 1. Topographical distribution of inflammatory or demyelinating lesions and of virus infected cells Region Inflamrnation/lesions Virus positive cells AE CPE SD E AE CPE SDE Lumbar spinal cord Thoracical spinal cord Cervical spinal cord Brain stem Cerebellar white matter Cerebellar cortex Basal ganglia Hippocampus Periventricular white matter Cerebral cortex Meninges 911 1 12/14 12/14